This research aims to assess the integration of different fuel cell (FC) options with battery and waste heat recovery systems through a mathematical modelling process to determine the most feasible retrofit solutions for a marine electricity generation plant. This paper distinguishes itself from existing literature by incorporating future cost projection scenarios involving variables such as carbon tax, fuel, and equipment prices. It assesses the environmental impact by including upstream emissions integrated with the Energy Efficiency Existing Ship Index (EEXI) and the Carbon Intensity Indicator (CII) calculations. Real-time data have been collected from a Kamsarmax vessel to build a hybrid marine power distribution plant model for simulating six system designs. A Multi-Criteria Decision Making (MCDM) methodology ranks the scenarios depending on environmental benefits, economic performance, and system space requirements. The findings demonstrate that the hybrid configurations, including solid oxide (SOFC) and proton exchange (PEMFC) FCs, achieve a deduction in equivalent CO2 of the plant up to 91.79% and decrease the EEXI and the average CII by 10.24% and 6.53%, respectively. Although SOFC-included configurations show slightly better economic performance and require less fuel capacity, the overall performance of PEMFC designs are ranked higher in MCDM analysis due to the higher power density.
Predictive maintenance (PdM) is vital to maritime operations; however, the traditional deep learning solutions currently offered heavily depend on centralized data aggregation, which is impractical under the limited connectivity, privacy concerns, and resource constraints found in maritime vessels. Federated Learning addresses privacy by training models locally, yet most FL methods assume homogeneous client architectures and exchange full model weights, leading to heavy communication overhead and sensitivity to system heterogeneity. To overcome these challenges, we introduce FLUID, a dynamic, model-agnostic FL framework that combines client clustering, structured pruning, and student–teacher knowledge distillation. FLUID first groups vessels into resource tiers and calibrates pruning strategies on the most capable client to determine optimal sparsity levels. In subsequent FL rounds, clients exchange logits over a small reference set, decoupling global aggregation from specific model architectures. We evaluate FLUID on a real-world heavy-fuel-oil purifier dataset under realistic heterogeneous deployment. With mixed pruning across clients, FLUID achieves a global of 0.9352, compared with 0.9757 for a centralized baseline. Predictive consistency also remains high for client-based data, with a mean per-client MAE of 0.02575 ± 0.0021 and a mean RMSE of 0.0419 ± 0.0036. These results demonstrate FLUID’s ability to deliver accurate, efficient, and privacy-preserving PdM in heterogeneous maritime fleets.
This study investigates the environmental and economic performance of integrating a proton exchange membrane fuel cell, battery systems, and an organic Rankine cycle-based waste heat recovery system for ship electrification. The analysis examines an onboard ammonia decomposition system for hydrogen production and ammonia production pathways. Additionally, the study benchmarks the effectiveness of onboard ammonia decomposition against green hydrogen bunkering scenarios (H2-BS). The analysis is based on data collected over two years from a bulk carrier provided by Laskaridis Shipping Co., Ltd. The environmental analysis includes well-to-wake emissions calculations. At the same time, economic performance is assessed through levelised cost of energy (LCOE) computations for 2025 and 2040, factoring in different fuel and carbon price scenarios. Consequently, the analysis utilises the Complex Proportional Assessment method to compare configurations featuring various ammonia production pathways across economic cases. The results indicate that green and pink ammonia feedstocks achieve maximum equivalent carbon dioxide reductions in the electrification plant of up to 47.28% and 48.47%, respectively, compared to H2-BS and 95.56% and 95.66% compared to the base scenario. Ammonia decomposition systems prove more economically viable than H2-BS due to lower storage and fuel costs, leading to competitive LCOE values that improve under higher carbon pricing scenarios.
Predictive maintenance for marine engines requires forecasts that are both accurate and technically interpretable. This work introduces BEACON, a hybrid architecture that combines a bidirectional long short-term memory encoder with attention pooling, a Kolmogorov Arnold network and a lightweight multilayer perceptron for cylinder-level exhaust gas temperature forecasting, evaluated in both centralized and federated learning settings. On operational data from a bulk carrier, BEACON outperformed strong state-of-the-art baselines, achieving an RMSE of , MAE of and of approximately 0.95, while producing interpretable response curves and stable SHAP rankings across engine load regimes. A second contribution is the explicit evaluation of explanation stability in a federated learning setting, where BEACON maintained competitive accuracy and attained mean Spearman correlations above between client-specific SHAP rankings, whereas baseline models exhibited substantially lower agreement. These results indicate that the proposed hybrid design provides an accurate and explanation-stable foundation for privacy-aware predictive maintenance of marine engines.
About Me
Innovation & Technology Manager in Laskaridis Shipping Co. LTD.
Innovation & Technology Manager in Laskaridis Shipping Co. LTD.